Welder multiphase circuit repair? direction question

[coppercone2]

My welding machine has 3x 20V rails. They seem to be regulated by zener diodes next to a HF transformer with 3 taps (3 rails).

One of the zener diodes, the same one, keeps blowing up on startup, even with the inverter board disconnected (no load).

If a transformer secondary has 3 taps, and there is some kind of over-voltage control problem in the primary, is it plausible that the same exact diode will break like 5/5 times on power up..? like the same tap keeps being effected. The other two diodes are fine on the other two taps.

Since it seems to be 3 identical unloaded circuits in parallel.. would the same exact circuit keep breaking every time? I figured it would have a chance to be any out of 3 on each power up. Is this normal in switchmode power circuits from a single transformer to keep blowing the same freaking tap every time? I replaced that diode more then a few times. My next plan was to 1-shot it, but I am kind of thinking there is only a few other parts on the PCB so I should just test the few power transistors.

[ESTechnical]

Can you work out enough of the circuit to make a schematic of the area surrounding the fault? Maybe post some photos of the board (both sides) too... It would probably help us to imagine the problem better.

Could the problem perhaps be related to the feedback in the switching supply, in such a way that the feedback is too slow to respond to a startup over-voltage? Do you have a scope to try and capture the turn on transient in the affected area? It might also help to be able to measure the current in this area.

[abdulbadii]

any device or PCB ID on which we'd google its schematic ?

[coppercone2]

No schematic, it just says 20V inverter rails.

I am just wondering about the transient analysis of a multi tap transformer. Why does it keep blowing the same diode. All the other component parts are fine, and the other 2 identical circuits are fine.

I figured it would randomly blow the first or third diode, but it always gets the third diode, since its an AC signal. I don't deal with these circuits much so I can be totally wrong, which is why I am asking. Usually I try to replace SwitchMode boards but this is expensive as hell

I.e. with this arrangement,

=====Primary=====
===============
=Sec1==Sec2==Sec3=

The circuit on circuit 3 always breaks, but its identical to sec1 and sec2, and none of them have a load. My scopemeter broke and I was kinda scared to try to trigger it on a non isolated scope.

I would figure.. either 1 or 3 would break, but its ALWAYS 3, I tried around 5 replacement diodes while I was fixing other crap.. so its a big sample size. I think it has to do with understanding correct transient analysis. I am not getting it,.

Their 22V zeners and MUR120 diodes next to each other, 3 sets.

Perhaps its time to fix that scope meter, people had good ideas on how to deal with it.

[ESTechnical]

Can you hand draw a schematic of the transformer, rectifier section and affected zener diodes? Photos of the top and bottom of the assembly would help too.

Do you have an isolation transformer and variac? We would probably try this PSU on its own, powered with isolated AC supply, so we could get a scope (with attenuator) on the affected part of the circuit.

Without knowing more about the circuit layout I'm a bit stuck as to why the same component would keep failing, but it';s totally possible some other component has failed and somehow destroying the zener with an overcurrent condition.

If you just wanted a measurement of the waveform on the transformer, you could possibly thread a thin insulated wire around the transformer windings to make one isolated turn to measure from. This would at least show you the shape of switching waveform during the start up, relatively safely on a scope. This should be enough to indicate if the startup is doing something crazy!